Comorbidity between verbal and non-verbal cognitive delays in 2-year-olds: a bivariate twin analysis

The purpose of this paper is to investigate the genetic and environmental aetiology of the comorbidity between verbal delay and non-verbal delay, in infancy. For more than 3000 pairs of 2-year-old twins born in England and Wales in 1994, we assessed verbal (vocabulary, V) and non-verbal (non-verbal, P) performance. V delay probands were selected who were in the lowest 5% of V; P delay probands from the lowest 5% of P. We assessed the comorbidity of delay both categorically, using twin cross-concordances, and dimensionally, by applying a bivariate extension of DeFries and Fulker (DF) group analysis. Both approaches are bidirectional, in that probands can be selected for either V delay (and analysed in relation to their co-twin's P score) or P delay (analysed in relation to their co-twin's V score). From a categorical perspective, twin cross-concordances indicated that comorbidity between V delay and P delay is substantially due to genetic factors whether probands are selected for V delay or for P delay. MZ and DZ cross-concordances were 24% and 8%, respectively, for probands selected for V delay and 27% and 6% for probands selected for P delay. From a dimensional perspective using bivariate DF analysis, selecting for V delay yielded high bivariate group heritability (0.59) and a genetic correlation of 1.0. In contrast, when selecting on P, DF analysis indicated lower bivariate group heritability (0.20) and only a modest genetic correlation with V assessed dimensionally (0.36). These results are discussed in terms of the difference between categorical and dimensional approaches to quantitative traits and the bidirectional nature of comorbidity. Such multivariate genetic results could lead to diagnostic systems that are based on causes rather than phenotypic descriptions of symptoms.